> we have no reason to believe their statements about risks are insincere

GPT-2 was too dangerous to be released.

We can argue about sincerity, but I don't think we can argue about utter historical incompetence in assessing the risks. It's one or the other.

Either way the evidence seems to indicate we should not listen to AI companies about the risks of AI. Which is not to say that there aren't risks, just that the dealer is the least credible review.

> GPT-2 was too dangerous to be released.

No, it was "let's set a precident while these things are not too dangerous, c'mon guys we know y'all can reproduce this easily".

GPT-2 was absolutely too dangerous to release at the time OpenAI made that statement. It’s only safe now because the specific risks they cited were dependent on the public’s lack of knowledge that such systems existed.